A coal sorting method, system, medium and equipment

Abstract

This invention discloses a coal sorting method, system, medium, and equipment, relating to the field of target detection technology. The method includes: acquiring real-time coal data from a mining area; performing target detection on the real-time coal data using an improved coal target detection model to obtain coal detection results; retaining or removing coal data from the coal detection results that meet preset confidence conditions according to a preset confidence interval; marking and storing the retained coal data; performing online training on the stored coal data through incremental learning to update the parameters of the improved coal target detection model; using the parameter-updated improved coal target detection model to perform coal target detection on newly acquired real-time coal data from the mining area to obtain real-time coal target detection results; and sorting the coal based on the real-time coal target detection results.

Description

Technical Field

[0001] This application relates to the field of target detection technology, and in particular to a coal sorting method, system, medium and equipment. Background Technology

[0002] The background of coal identification technology involves all aspects of coal mining, transportation, sorting, storage, processing and use. With the development of industrial automation and intelligence, coal identification technology has played an important role in improving the utilization efficiency and quality control of coal.

[0003] In existing technologies, traditional methods sort coal manually or mechanically, typically using sieving or crushing to classify coal types. This method is highly inefficient. Computer vision technology, which uses cameras to capture coal images and employs deep learning for automatic coal identification and classification, faces challenges. Existing deep learning object detection models exhibit biases in identifying coal from different mining areas or batches, leading to misclassification and reduced accuracy in coal target recognition.

[0004] Therefore, there is an urgent need for a coal sorting method to improve and ensure the accuracy of coal testing in different mining environments or different batches. Summary of the Invention

[0005] Therefore, it is necessary to provide a coal sorting method, system, medium, and equipment to address the aforementioned technical problems.

[0006] The present invention adopts the following technical solution:

[0007] This invention provides a coal sorting method, comprising:

[0008] Acquire real-time coal data from the mining area; use a coal target detection model to perform target detection on the real-time coal data from the mining area to obtain coal detection results; the coal detection results include: bounding box coordinates, confidence score, and category label;

[0009] Based on preset first, second, and third confidence intervals, coal data that meets preset confidence conditions in coal testing results is retained or removed, including: if the confidence level is within the first confidence interval, the corresponding coal data is removed; if the confidence level of the coal testing result is within the second confidence interval, the corresponding coal data is weighted and labeled, and the weighted coal data is retained and stored; if the confidence level of the coal testing result is within the third confidence interval, the corresponding coal data is retained and stored; the weighting and labeling of the corresponding coal data if the confidence level of the coal testing result is within the second confidence interval specifically includes: weighting and labeling the corresponding coal data within the second confidence interval... The coal data corresponding to the confidence level of the first confidence interval is labeled multiple times to obtain multiple labeling results for the coal data, including: bounding box coordinates, confidence level, and category label; the bounding box coordinates after multiple labeling are weighted to obtain weighted bounding box coordinates; the confidence levels of the same category label in the multiple labeling results are added together to obtain the category score, and the category label corresponding to the highest category score is the target category of the coal data corresponding to the confidence level of the second confidence interval, and the weighted bounding box coordinates and target category are retained; wherein, the maximum value of the first confidence interval is less than the minimum value of the second confidence interval; the maximum value of the second confidence interval is less than the minimum value of the third confidence interval;

[0010] By incremental learning, the stored coal data is trained online to obtain a coal target detection model with updated parameters.

[0011] The updated coal target detection model is used to detect coal in real time in the mining area, obtain the latest coal detection results, and sort the coal in real time based on the latest coal detection results.

[0012] Preferably, the coal target detection model is obtained by introducing a context-guided module into the YOLOv8 target detection model.

[0013] Preferably, the expression for the weighted bounding box coordinates is:

[0014] ;

[0015] In the formula, For the first The confidence level of the sub-label. , These are the four coordinate values ​​of the bounding box.

[0016] Preferably, the context-guided module includes: a convolutional layer, a local feature extractor, a surrounding context extractor, a joint feature extractor, and a global context extractor;

[0017] In the local feature extractor, local features of coal data processed by the convolutional layer are extracted;

[0018] The surrounding context extractor aggregates contextual information from coal data at different scales through spatial pyramid pooling.

[0019] In the joint feature extractor, the local features and the context information are concatenated, and the concatenated features are integrated through their batch normalized linear units and parameterized linear units to obtain the key features of the coal data;

[0020] In the global context extractor, context information is extracted from the key features of the coal data through global average pooling, and the context information is then processed in depth through two fully connected layers to obtain the final coal image feature map.

[0021] Preferably, the formula for extracting contextual information from key features of the coal data through global average pooling and performing in-depth processing of the contextual information through two fully connected layers is as follows:

[0022] ;

[0023] In the formula, This is the final feature map of the coal image. Features extracted by the feature extractor .

[0024] The present invention also provides a coal sorting system, comprising:

[0025] The data acquisition module is used to acquire real-time coal data from the mining area.

[0026] The target detection module is used to perform target detection on real-time coal data in the mining area using a coal target detection model to obtain coal detection results. Based on preset first confidence intervals, second confidence intervals, and third confidence intervals, it retains or removes coal data in the coal detection results that meet preset confidence conditions, including: if the confidence level is in the first confidence interval, then the corresponding coal data is removed; if the confidence level of the coal detection result is in the second confidence interval, then the corresponding coal data is weighted and labeled, and the weighted coal data is retained and stored.

[0027] The online training module is used to train the stored coal data online through incremental learning to obtain a coal target detection model with updated parameters.

[0028] The coal sorting module is used to perform target detection on real-time coal data in the mining area using a coal target detection model with updated parameters, obtain the latest coal detection results, and sort the coal in real time based on the latest coal detection results.

[0029] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described coal sorting method.

[0030] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the coal sorting method described above.

[0031] The above-mentioned at least one technical solution adopted in this invention can achieve the following beneficial effects:

[0032] The coal sorting method provided by this invention employs a coal target detection model with a context-guided module based on a preset confidence interval to perform target detection on the acquired coal data. Simultaneously, it captures and fuses local features, surrounding context information, and global context information of the coal data from the mining area, effectively improving the feature capture and learning capabilities of the YOLOv8 model and achieving high-precision coal detection. Based on the coal detection results, the coal data is filtered and weighted, resulting in high-quality coal data. Furthermore, by utilizing incremental learning and online training with high-quality coal data, a model operation mode of simultaneous detection and training is achieved.

[0033] In summary, based on the detection results obtained by the improved coal target detection model, this invention uses high-quality coal data corresponding to the detection results to train and update the coal target detection model in real time. This allows it to adapt to new coal data in mining areas in real time, ensuring the prediction accuracy of the coal target monitoring model and thus effectively guaranteeing the accuracy of coal target detection in different mining areas, different batches, and under different environmental changes. Attached Figure Description

[0034] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0035] Figure 1 This is a schematic flowchart of a coal sorting method provided by the present invention.

[0036] Figure 2 A schematic diagram of a multi-path convolutional and fusion neural network architecture for a coal sorting method provided by the present invention;

[0037] Figure 3 A schematic diagram of a multi-level feedback convolutional recurrent network for a coal sorting method provided by the present invention;

[0038] Figure 4A schematic diagram illustrating the continuous optimization and improvement process of a coal sorting method provided by this invention;

[0039] Figure 5 This is a schematic diagram of weighted labeling of coal data in a coal sorting method provided by the present invention;

[0040] Figure 6 A schematic diagram of the mechanical equipment for a coal sorting method provided by the present invention;

[0041] Figure 7 A schematic diagram illustrating the detection results of a coal sorting method provided by the present invention;

[0042] Figure 8 This invention provides a schematic diagram of a coal sorting system.

[0043] Figure 9 A schematic diagram of a computer device for implementing a coal sorting method provided by the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in the specification without creative effort are within the scope of protection of this application.

[0045] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0046] Figure 1 This is a schematic diagram of a coal sorting method according to the present invention, which specifically includes the following steps:

[0047] S101: Acquire real-time coal data from the mining area; use an improved coal target detection model to perform target detection on the real-time coal data from the mining area to obtain coal detection results; retain or remove coal data from the coal detection results that meet the preset confidence conditions according to the preset confidence interval; mark and store the retained coal data.

[0048] Optionally, coal-related images (such as mining areas, conveyor belts, coal piles, etc.) can be acquired through cameras, and the raw data can be cleaned (removing blur and duplicate images), normalized (unifying resolution), and data augmented (rotating, flipping, adding noise, etc.) to improve the model's generalization ability.

[0049] Optionally, the context-guided module includes: convolutional layers, local feature extractors, surrounding context extractors, joint feature extractors, and global context extractors.

[0050] Optionally, a local feature extractor is used to extract local features from the coal data processed by the convolutional layer; the surrounding context extractor aggregates contextual information of the coal data at different scales through spatial pyramid pooling.

[0051] Optionally, the joint feature extractor includes batch normalized linear units and parametric linear units, which are used to integrate the local features of the coal data and the features obtained by splicing the context information of the coal data at different scales through its batch normalized linear units and parametric linear units to obtain the key features of the coal data.

[0052] Optionally, the global context extractor extracts contextual information from key features of coal data through global average pooling, and performs deep processing on the contextual information through two fully connected layers to obtain the final coal image feature map.

[0053] Specifically, see Figure 2 The diagram illustrates a neural network architecture involving multi-path convolution and fusion, comprising a local feature extractor F1, a surrounding context extractor F2, a joint feature extractor F3, and a global context extractor F4. First, the input image, processed by standard convolutional layers, is then collaboratively processed by the local feature extractor F1 and the surrounding context extractor F2. The local feature extractor F1 focuses on the local features of the coal, while the surrounding context extractor F2 aggregates contextual information from different scales of the image through spatial pyramid pooling. This ensures that the model understands not only the information of each pixel but also the relationships between these regions within the overall context. Subsequently, the joint feature extractor F3, composed of batch normalization and parameterized rectified linear units, integrates the key coal information captured by the local feature extractor F1 and the surrounding context extractor F2. Finally, the global context extractor F4 extracts contextual information from the entire image through global average pooling and performs deep processing on this coal information through two fully connected layers to enhance the coal features learned by the joint feature extractor F3, thus obtaining the final output feature map features, as shown in the formula:

[0054] ;

[0055] In the formula, This is the final feature map of the coal image. Features extracted by feature extractor Fi .

[0056] Specifically, see Figure 3This diagram illustrates a multi-level feedback convolutional recurrent network. The context-guided module captures contextual features at all stages, fully utilizing information from both semantic and spatial levels. This is crucial for accurately classifying each pixel in an image. This paper improves YOLOv8 with the proposed context-guided module to obtain a better backbone feature extraction network. It simultaneously captures local features, surrounding context, and global context, fusing this information to improve the accuracy of semantic segmentation and enhance the feature capture and learning capabilities of the backbone network.

[0057] S102: Based on the preset first confidence interval, second confidence interval and third confidence interval, retain or remove coal data in the coal detection results that meet the preset confidence conditions;

[0058] Optionally, the maximum value of the first confidence interval is less than the minimum value of the second confidence interval; the maximum value of the second confidence interval is less than the minimum value of the third confidence interval.

[0059] Optionally, the preset confidence intervals include a first confidence interval, a second confidence interval, and a third confidence interval; the retention or removal of coal data that meets the preset confidence conditions in the coal testing results specifically includes: if the confidence level of the coal testing result is in the first confidence interval, then the corresponding coal data is removed; if the confidence level of the coal testing result is in the second confidence interval, then the corresponding data is weighted and retained; if the review is approved, the coal data is retained, otherwise, the coal data is removed; if the confidence level of the coal testing result is in the third confidence interval, then the corresponding data is retained.

[0060] Specifically, the preprocessed data is input into the improved YOLO-v8 model. The model analyzes the data, detects coal blocks as targets, attempts to detect coal targets within them, and outputs detection results, including the target's location, category, and confidence score. Based on the confidence score, three confidence intervals are defined: the first interval (confidence ≤ 50%): these are considered low-confidence predictions, classified as negative samples, and are considered incorrect detections; the second interval (50% < confidence < 90%): these are medium-confidence predictions, and the coal data corresponding to this confidence level are weighted and retained; the third interval (confidence ≥ 90%): these are high-confidence predictions, marked as correct predictions, and directly stored for training.

[0061] Optionally, if the confidence level of the coal detection result falls within the second confidence level interval, then the corresponding coal data is weighted and labeled. Specifically, this includes: labeling the coal data corresponding to the confidence level within the second confidence level interval multiple times to obtain multiple labeling results for that coal data, including: bounding box coordinates, confidence level, and category label; weighting the bounding box coordinates after multiple labelings to obtain weighted bounding box coordinates; and then... The weighted coordinates are represented as follows:

[0062] ;

[0063] In the formula, For the first The confidence level of the sub-label. , These are the four coordinate values ​​of the bounding box.

[0064] The confidence scores of the same category labels in multiple annotation results are added together to obtain the category score. The category label corresponding to the highest category score is the target category of the coal data corresponding to the confidence score of the second confidence interval. The weighted bounding box coordinates and target category are retained.

[0065] Specifically, the coal data corresponding to this confidence level is weighted and labeled. This includes: for coal data with a confidence level in the second confidence interval, weighted labeling is applied to data with confidence levels between 50% and 90% based on their bounding box position and category. The weighted labeled data is then stored in the incremental learning training dataset for incremental learning and online training. Different weights are assigned to data based on their confidence level. The higher the confidence level, the greater the weight allocated in the weighted labeling. For example, for a detection result with a confidence level of 0.8 and a data point with a confidence level of 0.6, 0.8 will have a higher weight in the weighted labeling. This is because a higher confidence level means the model is more confident in the detection result, thus giving it greater influence in the labeling process. (See [link to relevant documentation]). Figure 5 This is a diagram illustrating the weighted labeling of coal data. "coal" represents the coal target, and the numbers represent the weights after weighted labeling.

[0066] S103: The stored coal data is trained online through incremental learning to update the parameters of the improved coal target detection model.

[0067] Specifically, the model is incrementally learned and trained online using stored high-confidence labeled data. Incremental learning allows the model to continuously learn from new data based on existing knowledge, thereby continuously optimizing the model's parameters and performance.

[0068] Specifically, traditional object detection models are typically trained offline, making them ill-suited for dynamically changing real-world scenarios. This system, however, utilizes online learning to update the model in real-time with new, high-confidence data. This approach allows the model to continuously learn new features and patterns, improving the accuracy and adaptability of coal detection. Online learning enables the model to absorb new knowledge during operation without requiring large-scale offline retraining. This not only saves time and computational resources but also allows the model to adapt more quickly to changes in the environment and data distribution. For example, in coal production environments, the appearance and quality of coal may change with variations in the mining area; the online learning mechanism allows the model to capture these changes promptly, maintaining high detection accuracy. By filtering high-quality data through confidence levels and combining incremental learning with model updates, continuous model optimization and closed-loop operation are achieved.

[0069] S104: Utilize the improved coal target detection model with updated parameters to perform coal target detection on the newly acquired real-time coal data of the mining area, obtain real-time coal target detection results, and perform real-time sorting of coal based on the real-time coal target detection results.

[0070] Specifically, see Figure 4 The updated model can be reused to test newly collected data, thus forming a closed-loop continuous optimization process to achieve continuous optimization and improvement of the model.

[0071] Specifically, based on newly acquired labeled data (including positive and negative samples), the model undergoes online incremental training, dynamically updating model parameters to improve its adaptability to complex scenarios. The performance of the updated model is tested using a validation set or real-time data. Key metrics include: Precision: the proportion of correctly predicted targets in the detection results; Recall: the proportion of correctly detected actual targets; and mAP (mean accuracy): a comprehensive measure of detection accuracy. Evaluation results are categorized as follows: Performance Improvement: Confirming effective model optimization, proceeding to the model update stage; Performance Decrease: Reverting to the old model version and retraining; Deploying the new, performance-compliant model to the real-world application environment completes the iterative loop, continuously improving detection performance.

[0072] Additionally, this manual also provides information on mechanical equipment used for coal testing; see [link to documentation]. Figure 6 ,include:

[0073] Screw conveyor: Coal blocks are placed at the tail end of the conveyor, and then the screw conveyor pushes the coal forward along the closed tubular trough through the rotating screw blades. After the coal blocks are conveyed to the top of the machine, they fall freely onto the horizontal conveyor belt, which carries the coal blocks forward.

[0074] Horizontal conveyor belt: Coal is transported horizontally on a belt conveyor by a motor-driven conveyor belt, which uses friction to move the coal along the belt.

[0075] CCD Industrial Camera: A CCD industrial camera is installed above the conveyor belt to detect coal blocks. The CCD camera uses global shutter synchronous exposure and high-brightness light source to trigger shooting, ensuring no motion blur and achieving high-precision image acquisition. Then, utilizing the low-noise characteristics of CCD, the image is preprocessed, such as denoising, contrast enhancement, and geometric correction, to provide clear and stable input data for subsequent analysis.

[0076] The spray dust suppression device uses high-pressure fine water mist nozzles to convert water into fine water mist particles with a diameter of 20-50 micrometers. These particles actively adsorb airborne dust using their large surface area and surface tension. When the water mist collides with dust particles, the dust particles are wetted, agglomerated, and settle due to gravity, causing them to become heavier and detach from their suspended state, eventually settling onto the conveyor belt. The device uses a dust concentration sensor to monitor the particulate matter content in the environment in real time. An intelligent linkage control system dynamically adjusts the spray frequency and water volume, achieving "spraying when there is dust, stopping when there is no dust." This efficiently captures dust in the PM2.5 to PM10 range (with a capture efficiency of over 90%) while precisely controlling the surface humidity of the coal (typically with a moisture content increase of ≤1.5%), preventing coal agglomeration or conveyor belt slippage caused by excessive humidification.

[0077] High-performance computing: The high-performance computer primarily analyzes the detection results of coal blocks, performs inference on the input images, and outputs detection results including bounding box coordinates, category labels, and confidence scores. Data is analyzed and trained based on confidence levels. Low confidence scores (≤50%) are treated as negative samples; medium confidence scores (50% < < 90%) are weighted and labeled accordingly; high confidence scores (≥90%) are marked as correct predictions, directly stored, and used for training to form a new data model. The performance of the new model is then evaluated. If the new model's performance improves, confirming effective model optimization, the model is updated; otherwise, performance deteriorates, reverting to the old model version and retraining.

[0078] The detection result image shows the final detection phenomenon. (See attached image.) Figure 7 "coal" refers to coal, and the image shown is a result of a coal sorting method provided in this manual, which is detected by the mechanical equipment while the conveyor belt is moving.

[0079] In summary, by training while detecting, the model can adapt to new data distributions in real time, which is crucial for handling constantly changing environments (such as different lighting conditions, new target categories, etc.). Simultaneously, the model can continuously learn new patterns during use, improving detection accuracy and robustness, especially when facing previously unseen or rare objects. Traditional machine learning workflows require large amounts of labeled data to train the model. Training while detecting reduces reliance on pre-labeled data because the model can learn directly from real-time data. Furthermore, training while detecting utilizes computational resources more efficiently because it avoids frequently training the entire model from scratch, instead fine-tuning the existing model.

[0080] The above describes a coal sorting method provided by one or more embodiments of this specification. Based on the same idea, this specification also provides a corresponding coal sorting system, as shown in Figure 8.

[0081] Figure 8 A schematic diagram of a coal sorting system provided in this specification includes:

[0082] The data acquisition module 801 is used to acquire real-time coal data from the mining area.

[0083] The target detection module 802 is used to perform target detection on real-time coal data in the mining area using a coal target detection model to obtain coal detection results. Based on preset first confidence intervals, second confidence intervals, and third confidence intervals, it retains or removes coal data in the coal detection results that meet preset confidence conditions, including: if the confidence level is within the first confidence interval, then the corresponding coal data is removed; if the confidence level of the coal detection result is within the second confidence interval, then the corresponding coal data is weighted and labeled, and the weighted coal data is retained and stored.

[0084] The online training module 803 is used to train the stored coal data online through incremental learning to obtain a coal target detection model with updated parameters.

[0085] The coal sorting module 804 is used to perform target detection on real-time coal data in the mining area using a coal target detection model with updated parameters, obtain the latest coal detection results, and sort the coal in real time based on the latest coal detection results.

[0086] For specific limitations regarding a coal sorting system, please refer to the limitations regarding a coal sorting method described above, which will not be repeated here. Each module in the aforementioned coal sorting system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0087] This specification also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 A method for sorting coal is provided.

[0088] This instruction manual also provides Figure 9 The schematic diagram of the computer device shown is as follows: Figure 9 As shown, at the hardware level, this computer device includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then executes it to achieve the above. Figure 1 A method for sorting coal is provided.

[0089] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0090] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this invention.

Claims

1. A coal sorting method, characterized in that, include: Obtain real-time coal data from the mining area; A coal target detection model is used to detect coal targets in real-time mining area data and obtain coal detection results. The coal target detection model is obtained by introducing a context-guided module into the YOLOv8 target detection model; The coal detection results include: bounding box coordinates, confidence level, and category label; Based on preset first, second, and third confidence intervals, coal data that meets preset confidence conditions in coal testing results is retained or removed, including: if the confidence level is within the first confidence interval, the corresponding coal data is removed; if the confidence level of the coal testing result is within the second confidence interval, the corresponding coal data is weighted and labeled, and the weighted coal data is retained and stored; if the confidence level of the coal testing result is within the third confidence interval, the corresponding coal data is retained and stored; the weighting and labeling of the corresponding coal data if the confidence level of the coal testing result is within the second confidence interval specifically includes: weighting and labeling the corresponding coal data within the second confidence interval... The coal data corresponding to the confidence level of the first confidence interval is labeled multiple times to obtain multiple labeling results for the coal data, including: bounding box coordinates, confidence level, and category label; the bounding box coordinates after multiple labeling are weighted to obtain weighted bounding box coordinates; the confidence levels of the same category label in the multiple labeling results are added together to obtain the category score, and the category label corresponding to the highest category score is the target category of the coal data corresponding to the confidence level of the second confidence interval, and the weighted bounding box coordinates and target category are retained; wherein, the maximum value of the first confidence interval is less than the minimum value of the second confidence interval; the maximum value of the second confidence interval is less than the minimum value of the third confidence interval; By incremental learning, the stored coal data is trained online to obtain a coal target detection model with updated parameters. The updated coal target detection model is used to detect coal in real time in the mining area, obtain the latest coal detection results, and sort the coal in real time based on the latest coal detection results.

2. The coal sorting method as described in claim 1, characterized in that, The expression for the weighted bounding box coordinates is: ； In the formula, For the first The confidence level of the sub-label. , These are the four coordinate values ​​of the bounding box.

3. The coal sorting method as described in claim 1, characterized in that, The context guidance module includes: a convolutional layer, a local feature extractor, a surrounding context extractor, a joint feature extractor, and a global context extractor; In the local feature extractor, local features of coal data processed by the convolutional layer are extracted; The surrounding context extractor aggregates contextual information from coal data at different scales through spatial pyramid pooling. In the joint feature extractor, the local features and the context information are concatenated, and the concatenated features are integrated through their batch normalized linear units and parameterized linear units to obtain the key features of the coal data; In the global context extractor, context information is extracted from the key features of the coal data through global average pooling, and the context information is then processed in depth through two fully connected layers to obtain the final coal image feature map.

4. The coal sorting method as described in claim 3, characterized in that, The formula for extracting contextual information from key features of the coal data through global average pooling and then performing in-depth processing of the contextual information through two fully connected layers is as follows: ； In the formula, This is the final feature map of the coal image. Features extracted by the feature extractor .

5. A coal sorting system, characterized in that, include: The data acquisition module is used to acquire real-time coal data from the mining area. A coal target detection model is used to detect coal targets in real-time mining area data and obtain coal detection results. The coal target detection model is obtained by introducing a context-guided module into the YOLOv8 target detection model; The coal detection results include: bounding box coordinates, confidence level, and category label; The target detection module is used to retain or remove coal data that meets preset confidence conditions in coal detection results based on preset first confidence intervals, second confidence intervals, and third confidence intervals. This includes: if the confidence level is within the first confidence interval, removing the corresponding coal data; if the confidence level of the coal detection result is within the second confidence interval, weighting and labeling the corresponding coal data, and retaining and storing the weighted coal data; if the confidence level of the coal detection result is within the third confidence interval, retaining and storing the corresponding coal data. The weighting and labeling of the corresponding coal data if the confidence level of the coal detection result is within the second confidence interval specifically includes: […]. The coal data corresponding to the confidence level in the second confidence interval is labeled multiple times to obtain multiple labeling results for the coal data, including: bounding box coordinates, confidence level, and category label; the bounding box coordinates after multiple labeling are weighted to obtain weighted bounding box coordinates; the confidence levels of the same category label in the multiple labeling results are added together to obtain a category score, and the category label corresponding to the highest category score is the target category of the coal data corresponding to the confidence level in the second confidence interval, and the weighted bounding box coordinates and target category are retained; wherein, the maximum value of the first confidence interval is less than the minimum value of the second confidence interval; the maximum value of the second confidence interval is less than the minimum value of the third confidence interval; The online training module is used to train the stored coal data online through incremental learning to obtain a coal target detection model with updated parameters. The coal sorting module is used to perform target detection on real-time coal data in the mining area using a coal target detection model with updated parameters, obtain the latest coal detection results, and sort the coal in real time based on the latest coal detection results.

6. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of claims 1 to 4.

7. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in any one of claims 1 to 4.

Images